Artificial intelligence is now familiar to most Americans, but familiarity has not brought comfort. A new national survey finds a public that is far more uneasy than excited about what AI could mean for the country in the next decade, with majorities across party lines saying government has not done enough to regulate it.
The poll, conducted by the Annenberg Public Policy Center of the University of Pennsylvania, comes at a moment when the country is arguing over how quickly to build the infrastructure behind AI, who should set the rules, and how much disruption the technology may bring to daily life. The answers suggest that many people are already paying attention, and a large share do not like where things appear to be heading.
Only 17% of Americans said AI would have a somewhat or very positive impact on the United States over the next 10 years. Far more, 42%, said the effects would be somewhat or very negative. Another 32% said the technology’s impact would be equally positive and negative.
That gap matters because awareness is already high. Nearly 8 in 10 Americans, 78%, said they had heard at least a moderate amount about AI, and 67% said they had used it at least a few times in the previous month.
Shawn Patterson Jr., a research analyst at the Annenberg Public Policy Center, said the findings point to a public that is engaged and increasingly worried. “These results tell a clear story: Americans are paying attention to AI, nearly 8 in 10 say they’ve heard at least a moderate amount about it, and what they see concerns them,” he said. “The demand for regulation is not a partisan issue. Majorities across the political spectrum say the government has done too little.”
One bright spot in a darker picture
Americans did not judge every possible use of AI the same way.
Medical research and discoveries stood out as the one area where optimism clearly outweighed concern. In that category, 57% of respondents said they expected AI to have a somewhat or very positive impact. No other area came close.
Elsewhere, confidence fell sharply. Just 24% expected a positive effect on government effectiveness. Only 22% said the same for the creative arts, and 19% for the economy. Views were even bleaker on questions tied more directly to everyday stress and instability. Only 17% expected AI to help mental health and well-being, 14% thought it would have a positive effect on household utility costs, and just 5% saw a positive impact on U.S.-China relations.
The pattern suggests that many Americans can picture AI helping scientists make discoveries, but they are much less convinced it will improve institutions, lower costs, or make life feel steadier.
That skepticism also appears in how people think about the technology overall. Asked to look ahead 10 years, only 7% described AI’s future impact on the United States as very positive, while 22% called it very negative. Another 11% said somewhat positive, and 20% said somewhat negative.
Regulation draws support from all sides
The strongest point of agreement in the survey involved regulation.
Nearly two-thirds of Americans, 65%, said the government has done too little to regulate AI. Only 8% said it has done too much, while 26% said the current level is about right. Support for stronger government action appeared in every major political group, including 77% of Democrats, 72% of independents, and 53% of Republicans.
A slim majority, 52%, said the federal government should take the lead rather than the states.
Concern rose along with pessimism, but it did not depend entirely on it. Among those who said AI’s impact would be very negative, 83% said the government has done too little. Even among those who expected AI to be very positive, 43% still wanted more regulation. A majority of people who thought AI would bring an equal mix of benefits and harms, 57%, also said the government had fallen short.
Only one group broke from that pattern: people who believed AI’s effects would be very positive. In that small group, 53% said regulation was about the right amount.
Matt Levendusky, a University of Pennsylvania professor and the Stephen and Mary Baran Chair in the Institutions of Democracy at APPC, said that relative consensus is unusual in a deeply divided political climate. “The relative lack of partisan polarization on AI is particularly striking in this political moment, given the ongoing political polarization on almost all other issues,” he said. “Concern about AI is bipartisan, and the public is waiting to see what politicians will do. This offers real potential rewards to either party if they can convince the public that they have the correct approach.”
A political issue, but not yet a fully partisan one
That unusual mix showed up elsewhere in the findings.
When respondents were asked whether Kamala Harris would have done a better or worse job than President Donald Trump across eight policy areas, AI regulation produced more overlap between the parties than other topics. It drew the highest share of “about the same” responses, 24%, of any issue tested.
That does not mean there were no partisan differences. There were still large gaps between Democrats and Republicans over who would handle AI regulation better. But the split was narrower than on immigration, inflation, or the economy. On AI regulation, 22% of both Democrats and Republicans said Harris would have done about the same as Trump. Regarding immigration, only 7% of Democrats and 6% of Republicans said that. As to the economy, the figures were 11% and 10%. On inflation, 12% and 11%.
The overall partisan gap was smaller as well. There was a 65-point difference between Democrats who said Harris would do better on AI regulation, 73%, and Republicans who said the same, 8%. That was still wide, but it was below the 75- and 76-point gaps seen on immigration, inflation, and the economy.
For a subject now tied to jobs, infrastructure, and national competitiveness, AI may still be new enough that fixed party lines have not completely formed.
The machines behind AI face their own backlash
The public’s unease extends beyond software and into the physical systems that power it.
As data centers spread across the country, often bringing worries about electricity demand and local disruption, 49% of Americans said they oppose building new data centers in their area. Nearly a third, 31%, said they strongly oppose them. Only 21% said they support new construction, while 30% were neutral.
There is also anxiety about work.
Among people who are employed, 41% said they were somewhat or very worried about losing their job or seeing their hours reduced because of AI. Democrats reported the highest level of concern at 50%, followed by independents at 41% and Republicans at 32%.
The survey was conducted among 1,330 U.S. adult citizens from Feb. 17 to March 20, 2026. It was carried out for the Annenberg Public Policy Center by SSRS, primarily online with a small sample of phone respondents. The margin of error for the full sample was plus or minus 3.5 percentage points.
Practical implications of the research
The findings suggest that public concern over AI is no longer limited to abstract fears about the future. Americans are connecting the technology to medical breakthroughs, job security, local development, energy use, and government responsibility.
That gives lawmakers, regulators, and companies a more complicated landscape to navigate. People appear open to AI’s benefits in medicine, but they are wary of broader social costs and want clearer rules.
Any effort to expand AI systems, or the data centers that support them, is likely to face a public that is informed, skeptical, and asking tougher questions than the industry may have expected.
- How artificial intelligence can reduce selfish behavior and reshape society
- Artificial intelligence can now beat average humans in creativity, study finds
- Women think artificial intelligence is riskier than men do, study finds